Review comments on draft-ietf-shim6-failure-detection-03.txt

[Geoff Huston <gih@apnic.net>]

With my WG Chair hat OFF, i'd like to offer some comments on the failure 
detection draft

Overall I found this to be a clear specification  - I have two comments / 
queries on the draft

1. Exploration termination

   Its clear that once you get bidirectional probe messages being 
exchanged the exploration algorithm terminates sucessfully with new locators.

   But what condition sets a no viable locator found state?

   Section 6.4 includes the comment that garbage collection on contexts 
terminates contexts that
   have failed due to the inability of the exploration process to find a 
working pair.

    section 5.3 is unclear as to termination - all I see is the text: 
"retrying the process as is necessary"

  Is there a termination condition for exploration?

2.  Probe timing

   What is the timing base for sending probe packets? I note that 5.3 
indicates that successive probe packets should undertake exponential 
backoff, but is there a start timer interval for probe packets?

    It seems to me that a relatively short initial timer can be used, and 
the exponential backoff  would extend this to longer inter-probe timings 
pretty rapidly.

   On a related note, is there a suggested maximum inter-probe time? At 
what point would exponential backoff not be applied any further?


Also the side note section 6.4 uses the term "happy" - is this a 
measureable  term? ;-)

regards,

   Geoff